Illustrator MCP

    1

    Let your AI agent edit a real, open Adobe Illustrator document: layers, text, vectors, export. Not an image generator.

    Free

    8 usesSecurity scanned

    Works with the AI Agents you already use

    Claude CodeClaude CodeCursorCursorCodex CLICodex CLIMuseMuseDotsDots+21 more

    Illustrator MCP

    Example session with this skill installed

    Vectorize ~/Desktop/client-logo.png into clean paths, center it on a new 1080x1080 artboard, and export a 2x transparent PNG to my Desktop.

    • Read your context and instructions
    • Compiled the illustrator mcp

    The agent first calls illustrator_get_status to confirm Illustrator is running and the server is connected. It then creates a 1080x1080 RGB document with illustrator_create_document and runs illustrator_vectorize_image on the PNG with the default VTracer engine. If VTracer isn't installed, it says so and retries with Illustrator's built-in Image Trace instead of failing.

    The traced logo lands in the document as real compound paths, not an embedded image. The agent selects it, centers it horizontally and vertically on the artboard with illustrator_align_selection, and reports how many paths were created.

    Finally it calls illustrator_export_document with format png, scale 200 and transparent true, saves the file to ~/Desktop, and replies with the exact output path. The .ai document stays open, so you can tweak anchor points, colors or layers by hand exactly as you normally would.

    Connects securely to your tools. The creator never sees your data.

    About this skill

    Most "AI design" tools hand you a flat picture and leave you to redo it by hand. This skill teaches your agent to work inside a real Adobe Illustrator document instead: the one open on your screen, with its layers, artboards, live text and vector paths, all still editable after the agent is done.

    It's the companion skill for Illustrator MCP, a free, open source (MIT) MCP server that connects Claude, Cursor or ChatGPT to Illustrator through its own scripting engine. The server gives your agent 29 tools. This skill tells the agent when to use which one, the coordinate and color conventions, step-by-step recipes for common jobs, and how to handle the risky ones safely.

    What your agent can do with it:

    • Set up a document in seconds: "Create a 1080x1080 RGB document with the title SALE in bold orange" gives you a real artboard with live text.
    • Vectorize a raster logo into editable paths, with VTracer or Illustrator's built-in Image Trace as fallback.
    • Draw, align and group shapes: "Draw six circles in two rows and center them on the artboard."
    • Recolor, move, rotate, scale and reorder whatever is selected, and organize it in layers.
    • Export PNG, JPG, SVG or PDF, including 2x transparent PNGs.
    • Run custom ExtendScript for anything the built-in tools don't cover. The skill tells the agent to explain the script and ask before anything destructive.

    Requirements: Adobe Illustrator on macOS or Windows, Node.js 18+, and the Illustrator MCP server installed and connected to your client. Setup takes a few minutes and is documented at https://github.com/gherardo200-glitch/illustrator-mcp

    Everything runs locally against your own copy of Illustrator. No files leave your machine.

    • Needs Adobe Illustrator installed and a local MCP server; it does not work in a browser-only or cloud-only agent.
    • It is not an image generator: it edits real Illustrator documents.
    • On macOS, Illustrator asks for Automation permission on first use.
    • Text and shape tools cover common cases; complex effects, gradients or brushes need run_script.

    How to use

    1. 1

      Connect your agent to the Agensi MCP

      Copy the prompt and paste it into Claude, ChatGPT, Cursor or any agent that supports MCP. It sets up the connection for you. Manual setup

    2. 2

      Sign in to Agensi

      Your agent shows a sign-in link. Sign in and click Allow.

    3. 3

      Ask for what you need

      Your agent loads this skill and uses it on your task. It always asks before any payment.

    Reviews

    No reviews yet

    Be one of the first to try it. Every listed skill goes through the trust checks below.

    Security scanned

    Checked by our 8-point scan before listing

    8 uses

    Times used to date

    Free forever

    No account required to browse

    Trust & safety

    Security scan

    Security scan passed 1 month ago

    • Free with an Agensi account

    Requires

    Adobe Illustrator
    macOS
    Windows
    Node.js 18+
    Illustrator MCP
    Node.js

    Permissions

    Read Files
    Write Files
    Network Access

    The skill itself contains only instructions, no scripts. All actions happen through the Illustrator MCP server running locally. The illustrator_run_script tool can execute arbitrary ExtendScript; the skill tells the agent to explain the script and ask before anything destructive.

    Listed1 month ago

    What's inside

    Frequently Asked Questions

    You need Adobe Illustrator on macOS or Windows, Node.js 18+, and the free Illustrator MCP server (MIT) installed and connected to your MCP client, such as Claude Code, Claude Desktop or Cursor. The server is not included in this skill's ZIP: setup instructions are at https://github.com/gherardo200-glitch/illustrator-mcp

    Unlike standard AI image generators that produce flat pixels (JPEGs/PNGs), this skill creates actual vector paths, layers, and anchor points inside Illustrator. This allows you to manually edit every curve, change colors, and scale artwork without losing quality.

    Yes, the 'image vectorization' feature allows you to import existing raster images and convert them into clean, editable vector paths using integrated tracing engines directly within your workspace.

    The skill includes a built-in coordinate conversion system that translates natural language requests into precise mathematical placements, handling the complex Y-axis orientation and color mode mapping (RGB vs. CMYK) automatically.

    Yes. The agent can select objects by name (or select everything), then recolor, move, rotate, scale or reorder the selection in one go, for example recoloring elements to match a new brand palette. For anything the built-in tools don't cover, it can run a custom ExtendScript, explaining it and asking before anything destructive.

    New in Branding & Design